There are limits to how good of a prediction we can make, and often they’re quire low.
We often under appreciate how low the ceiling is in terms of our ability to predict the future. Doesn’t matter how much machine learning and personal resources you throw into predicting the next roll of a fair die… we can’t get better than 1/6th accuracy.
Expert predictors are often grossly wrong about the level of objective ignorance that exists in their domain of expertise.